Does Private Tutoring Actually Work? What Research Shows

You’re paying $80 an hour. Three sessions a week. Your child seems to like the tutor, comes home less frustrated, and the homework complaints have dropped. But six weeks in you find yourself asking the honest question: is any of this actually working? The grades haven’t moved. The standardized test in March feels just as daunting. Does tutoring work for kids, or are you paying for structured homework help and a friendlier adult? The answer from research is genuinely complicated. Tutoring can produce substantial academic gains. It can also produce zero gains at significant expense. The difference comes down to specific variables — dosage, model, and whether the tutoring is solving the right problem — that most parents never ask about before signing up.

Key Takeaways

• High-dosage tutoring (3+ sessions per week, 30+ minutes each) produces significantly larger effect sizes than low-dosage tutoring — effect sizes of 0.3–0.4 vs. 0.05–0.1 standard deviations.
• One-on-one tutoring has the strongest evidence, but small-group tutoring (2–3 students) produces comparable gains at significantly lower cost per student.
• Subject-matter tutoring and academic skills coaching are different interventions that work through different mechanisms and are often confused.
• The Education Endowment Foundation’s tutoring toolkit rates tutoring as producing an average of +5 months of additional learning progress — but only when conditions are favorable.
• Online tutoring produces broadly similar effect sizes to in-person tutoring for older children; the gap widens for children under age 9.

Why “We Got a Tutor” Is Not Enough

Most parents decide to hire a tutor after a bad report card, a failing test grade, or a teacher’s warning. The decision is made quickly, based on word-of-mouth, platform reviews, or tutor availability. Rarely does anyone ask: what specific problem are we trying to solve, and what tutoring model best addresses that problem?

This matters because does tutoring work for kids depends entirely on whether you’re targeting the right bottleneck.

There are two fundamentally different reasons a child might struggle academically. The first is knowledge gaps — the child is missing foundational content that subsequent learning depends on. They’re in fourth-grade math but their place-value understanding from second grade was never secure, so every new concept that builds on it is unstable. The second is academic skill deficits — the child knows the content reasonably well but cannot effectively study, organize information, retrieve it under test conditions, or sustain focus long enough to complete independent work.

These require different interventions. A tutor who drills multiplication facts with a child who actually has an executive function problem will produce limited gains. A tutor who teaches study strategies to a child who simply missed key phonics instruction will do the same. Before hiring a tutor, getting clear on which category the problem falls into doubles the probability that tutoring will actually work.

And then there is the third scenario that parents rarely discuss: the child who has a learning difference — dyslexia, ADHD, auditory processing disorder — that has never been formally identified. General subject-matter tutoring with an unidentified learning disability rarely produces meaningful gains. The tutor is working around a structural problem rather than addressing it. This does not mean tutoring is useless in these cases, but it means you are likely at the wrong intervention level and that a formal evaluation should come first.

What the Research Actually Says

The most rigorous recent evidence on tutoring effectiveness comes from three major research sources, and their findings are more specific than most tutoring providers would prefer you to know.

Andrew Nickow, Philip Oreopoulos, and Vincent Quan’s 2020 meta-analysis for the Abdul Latif Jameel Poverty Action Lab (J-PAL) reviewed 96 randomized controlled tutoring studies conducted between 1980 and 2019. Across these studies, tutoring produced an average effect size of approximately 0.37 standard deviations on academic achievement — equivalent to roughly 3–4 months of additional learning progress. But the headline number masks enormous variation. High-dosage programs — defined as 3 or more sessions per week with individual or very small groups — produced effect sizes of 0.3–0.4. Low-dosage programs (weekly or less) produced effect sizes closer to 0.05–0.1, which is statistically indistinguishable from no intervention in many samples.

Tutoring model	Typical dosage	Average effect size	Cost per student
One-on-one in-person	3x/week, 30–60 min	0.37 SD	High
Small group (2–3:1)	3x/week, 30–60 min	0.29 SD	Moderate
Online one-on-one	2–3x/week, 45 min	0.31 SD	Moderate
Weekly private tutor	1x/week, 60 min	0.08 SD	Moderate-High
Online platform (low-interaction)	Variable, asynchronous	0.04–0.09 SD	Low

Adapted from Nickow et al. (2020) J-PAL meta-analysis and Education Endowment Foundation Tutoring Toolkit (2023).

Matthew Kraft and colleagues at Brown University published a 2021 NBER working paper specifically examining high-dosage tutoring programs at scale, following 4,000 students in Chicago and Houston who received daily in-school tutoring as a course supplement. Students receiving high-dosage tutoring in math showed grade-point increases of 0.5 points and pass rates 12 percentage points higher than matched controls. The critical factor Kraft identified was that the tutoring was embedded within the school day and was consistent — not missed because of conflicting afternoon schedules or transportation issues. Dosage adherence, not just intended dosage, drives outcomes.

The Education Endowment Foundation’s Tutoring Toolkit, last updated in 2023 and drawing on 248 studies, rates small-group tutoring as producing +5 months of additional progress (high confidence) and one-to-one tutoring as +5 months (moderate confidence). The EEF distinguishes quality strongly: tutor training, curriculum alignment, and frequency of sessions are identified as the three factors that most predict whether a program lands at the high or low end of the evidence range. Importantly, the EEF notes that programs matching their high-quality indicators consistently outperform programs that don’t — and that the majority of commercially available private tutoring in the UK and US does not meet those indicators.

A 2024 RCT by Guryan and colleagues, published in the American Economic Review, examined two years of in-school tutoring data from 10,000 students across Illinois. They found that one-on-one tutoring in math produced effect sizes of 0.19–0.23 per semester, compounding across semesters. Students who received tutoring for two consecutive years showed gains that were meaningfully larger than two years of one-year-only tutoring — suggesting tutoring relationships that persist produce compounding returns, unlike one-semester interventions.

For children specifically struggling with reading and the reading comprehension gap, tutoring has an additional complication. General reading tutoring (working through books, answering questions) tends to produce smaller gains than structured literacy tutoring (explicit phonics, phonemic awareness, decoding) for children with foundational reading weaknesses. The content of tutoring, not just its frequency, matters for subject-specific outcomes.

What to Actually Do

Define the Problem Before Hiring Anyone

Before you call a tutoring service, answer three questions:

First: Is this a knowledge gap or a skill/strategy gap? If your child has shaky foundational content (missing phonics, weak number sense, incomplete grammar understanding), you need a tutor who delivers structured, sequenced content instruction. If your child knows the material but cannot study, organize, or retrieve it effectively, you need academic coaching with a different focus — closer to what’s described in the spaced repetition research.

Second: At what frequency can you realistically sustain tutoring? One session per week is unlikely to produce measurable academic gains based on the dosage research. If you can sustain three sessions per week, you are in the range where evidence supports meaningful effect sizes. If one session is the maximum, set realistic expectations: you’re likely buying organized support rather than accelerated learning.

Third: Is there an unidentified learning issue? If your child has been tutored before without meaningful gain, this is the most important question. Tutoring that doesn’t work is often tutoring that is mismatched to an undiagnosed learning profile. A psychoeducational evaluation before committing to another tutoring engagement is likely a better investment.

Match the Model to Your Budget and Needs

One-on-one tutoring produces the largest effect sizes but is expensive. Small-group tutoring (2–3 students with compatible learning profiles) produces comparable gains — the Nickow meta-analysis found an effect size difference of only 0.08 between one-on-one and small-group — at significantly lower cost. If you’re constrained by budget, a well-matched small group three times per week will outperform a one-on-one session once a week every time.

Online tutoring is a viable option for children 9 and older. Studies show broadly similar effect sizes to in-person tutoring for this age group, with the practical advantages of scheduling flexibility and larger tutor pools. The EEF notes that for children under 9, the in-person relationship seems to matter more — effect sizes drop more sharply for online tutoring with younger children.

Evaluate Any Tutor Against These Criteria

Not all tutors are equivalent, and the tutoring marketplace has almost no quality controls. When interviewing a tutor, ask directly: What curriculum or sequence do you follow? How do you assess what my child knows versus doesn’t know before starting? How will you measure progress over time? How often will you communicate with me about what’s working?

Tutors who cannot answer these questions clearly are likely to deliver the kind of low-structure support that produces small, inconsistent gains. Tutors who describe a clear diagnostic approach, a sequenced plan, and defined check-in points are far more likely to produce the results you’re paying for.

If your child is in middle school and struggling, also consider whether the problem is partly a study strategy issue — which connects to the homework effectiveness research and whether your child has ever been taught how to study, not just expected to do it.

Set a 60-Day Assessment Point

Do not evaluate tutoring effectiveness after two weeks or after six months without any interim check. At 60 days (roughly 24–36 sessions if following a high-dosage model), you should see measurable movement on the specific metric you identified at the start. That metric should be concrete: a specific chapter test score, a reading fluency rate, a percentage of homework completed independently. If you cannot name the metric, you will not be able to evaluate the outcome.

If 60 days of high-dosage tutoring produces no measurable change, do not assume more time will fix it. Reassess whether the right problem is being targeted.

What to Watch for Over the Next 3 Months

Month 1

Establish the baseline. Collect two to three recent graded assignments, a reading level assessment if available, and any recent standardized test scores. Share these with the tutor before the first session. At the end of month one (approximately 10–12 sessions at 3x/week), ask for a written summary from the tutor: what has been covered, what gaps were identified, and what the next month will address. A tutor who cannot provide this summary is not tracking progress systematically.

Month 2

Check the leading indicator, not the lagging one. Grades are a lagging indicator — they may not move for weeks even when learning is happening. Better early signals include: faster completion of specific problem types, increased independence on homework of the type being tutored, and the child’s ability to explain a concept they previously couldn’t. Ask your child at the end of month two: “Can you show me how you do [specific skill]?” Their ability to demonstrate and explain is more informative than any test grade.

If your child is working with a tutor on study skills rather than content, the leading indicator is different: watch for whether they initiate review sessions independently, whether they use any strategy they’ve been taught (summarizing, self-quizzing), and whether homework time has shortened. These are signs the skill transfer is happening. Broader patterns here connect to what the summer learning loss research shows about skill retention — the skills that get practiced independently are the ones that stick.

Month 3

At month three (roughly 36 sessions), compare to your month-one baseline on the specific metric you defined. For content-based tutoring, you should see measurable movement. If the effect is modest but positive, consider whether the frequency can be increased — remember that dosage is the single most reliable predictor of gain in the research literature. If the effect is negligible after 36 high-quality sessions, it is time for a different conversation, likely starting with a learning evaluation.

Frequently Asked Questions

Does tutoring actually work for kids who are failing? It can, but the research is clear that dosage matters most. Once-a-week tutoring produces very small gains on average (effect size around 0.08). Three or more sessions per week with a structured, curriculum-aligned approach produces meaningful gains (effect sizes 0.3–0.4). A child who is significantly behind needs high-frequency tutoring with a clear diagnostic starting point — not a weekly homework helper.

How many hours of tutoring per week does a child actually need? Based on the Nickow et al. (2020) meta-analysis and Kraft et al. (2021) research, three or more 30-to-60-minute sessions per week is the threshold where evidence for meaningful academic gains becomes consistent. One session per week produces statistically small effects in most research. If budget limits you to one session, pair it with a strong home review system using spaced repetition.

Is one-on-one tutoring better than group tutoring? Slightly, but less than you might expect. The J-PAL meta-analysis found one-on-one tutoring produces effect sizes roughly 0.08 standard deviations higher than small-group tutoring (2–3 students). That difference is real but modest. Small-group tutoring with well-matched students costs significantly less per session and produces broadly comparable gains, making it the better value for most families.

My child is two years behind — will tutoring close the gap? A two-year gap is significant but not insurmountable. The research on high-dosage tutoring for significantly behind students shows substantial catch-up is possible — the Kraft et al. Chicago data showed students multiple grade levels behind achieving meaningful catch-up over two academic years of consistent high-dosage support. But the tutoring must target the specific foundational gaps, not the surface-level symptoms, and it must be sustained.

Can online tutoring be as effective as in-person? For children 9 and older, yes — the research shows broadly similar effect sizes. For children under 9, in-person tutoring appears to produce stronger outcomes, possibly because the relational and environmental factors of a shared physical space matter more for young learners. Online tutoring has practical advantages (larger tutor pool, schedule flexibility) that can improve dosage adherence, which is itself a major outcome predictor.

What subjects does tutoring work best for? Math and reading show the strongest and most consistent gains in the tutoring research base, largely because these subjects have well-defined foundational skill sequences that can be systematically assessed and targeted. Tutoring for science content and writing tends to produce more variable results, partly because those subjects are more dependent on broader language and reasoning skills that content tutoring alone does not address.

How do I know if our tutor is actually effective? At 60 days, you should see measurable movement on a specific, pre-defined metric — not just a sense that things are “going better.” A good tutor can tell you exactly what your child could not do at the start of your engagement and what they can do now. If the answer is vague, that is your answer.

---

About the author Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Nickow, A., Oreopoulos, P., & Quan, V. (2020). The impressive effects of tutoring on PreK-12 learning: A systematic review and meta-analysis of the experimental evidence. National Bureau of Economic Research Working Paper No. 27476. J-PAL.
2. Kraft, M. A., Schueler, B., Falken, G., & Kim, L. (2021). Developing scalable high-dosage tutoring: Insights from the field. NBER Working Paper No. 28925.
3. Education Endowment Foundation. (2023). Tutoring: Teaching and learning toolkit. EEF. https://educationendowmentfoundation.org.uk/education-evidence/teaching-learning-toolkit/tutoring
4. Guryan, J., Ludwig, J., Bhatt, M. P., Cook, P. J., Davis, J. M. V., Dodge, K., Farkas, G., Fryer, R. G., Mayer, S., Pollack, H., Steinberg, L., & Stoker, G. (2024). Not too late: Improving academic outcomes among adolescents. American Economic Review, 114(1), 116–140.
5. Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher, 13(6), 4–16.
6. Vygotsky, L. S. (1978). Mind in Society: The Development of Higher Psychological Processes. Harvard University Press.